10eastern.
boards / oekaki / #22457

by fatribz, idwtwwr

Comments (6)

fatribz Mar 29, 2009 · 12:56 PM
i don't understand what is going wrong
fatribz Mar 29, 2009 · 5:00 PM
pixels are fuct. somtimes it works, sometimes it fails. i can't reproduce the shit. how would you turn a full color its its n most distinct colors? how do you map that shit? because my tolerance technique doesn't do shit. horizontal lines.. retarded misscolored stuff. did i miscalculate? am i gonna get retarded again? tuff stuff everywhere.
yanbu Mar 29, 2009 · 7:43 PM
some tough questions. i wish i had answers, but this lookss great
allman vibe Mar 29, 2009 · 10:28 PM
dang! what!
ds Mar 29, 2009 · 10:37 PM
ribz, i too fell in to the color quantization pit for similar reasons. the median cut algorithm is popular, as are octrees. use one of those to get your N colors, then the shortest geometric distance can between the table color and the source color is the best match. i think...i've not had much time to play with it yet.
fatribz Mar 30, 2009 · 2:29 PM
thanks a lot for the heads up! i didn't know what to even search for